Hazard Analysis

Purpose and output

Hazard analysis is the reasoning core of HACCP. It identifies the hazards reasonably expected at each step, evaluates which are significant, and determines how they will be controlled. Its output is not merely a table of organism names. It is a documented explanation linking product, process, intended use, hazard behaviour, evidence and control. The analysis must be completed by people with the necessary product, process and technical knowledge and must describe the operation that actually exists. A hazard omitted at this stage cannot be managed coherently later, while a hazard included without meaningful evaluation can produce unnecessary critical control points and unusable monitoring. The analysis should be precise enough to guide decisions yet capable of review when knowledge or operating conditions change.

Define product, scope and intended use

The team first establishes the boundaries of the analysis. The product description covers species, raw materials, formulation, curing agents, allergens, physical form, casing or package, intrinsic characteristics, processing, storage, shelf life and distribution. Intended use states whether the food is ready to eat or will receive a further validated treatment, and whether vulnerable consumers are expected. A product sold as ready to eat cannot assign its final lethality duty to the consumer. The analysis also identifies activities inside and outside the establishment, such as supplier controls, outsourced processing and transport. Similar products may be grouped only when their hazards and controls are genuinely equivalent. A change in diameter, fat content, starter system or packaging can be enough to break that equivalence.

Map and confirm the process

A flow diagram records every relevant step from receiving to distribution. For cured meat this can include storage, thawing, trimming, grinding, mixing, formulation, curing, equilibration, stuffing, fermentation, smoking, heating, cooling, drying, maturation, slicing, packaging, finished storage and dispatch. Rework, delays, temporary holds, cleaning transitions and product returned to an earlier stage must be included where they occur. The team confirms the flow on site and across different shifts or production patterns. This verification often reveals conditions absent from written procedures, such as warm staging before fermentation, probe placement that misses the slowest unit, condensation above exposed ready-to-eat product or mixed use of equipment. Hazard analysis follows the real flow; otherwise the controls can be internally logical and operationally false.

Identify hazards at every step

For each ingredient and process step, the team asks which biological, chemical and physical hazards may be present or introduced, whether an existing hazard may increase, whether it may survive a treatment, and whether a controlled product may be recontaminated. Allergen and label hazards are added according to the legal and system framework. Sources include animals and raw meat, water, spices and cultures, employees, equipment, the environment, packaging and storage. The analysis should be specific: Salmonella survival during a named non-thermal process is more useful than the word bacteria; excessive nitrite caused by a weighing or mixing error is more useful than chemicals. Specificity allows the team to select a measure and define what evidence must support it.

Evaluate severity and likelihood

The team evaluates the severity of harm and the likelihood of occurrence in the absence of control. Likelihood is not the same as frequency in finished-product test results. It considers prevalence in raw materials, growth or survival potential, process history, environmental exposure, supplier and establishment data, scientific literature, epidemiology, intended use and susceptible consumers. Severe hazards may require control despite infrequent occurrence. Conversely, a hazard that is biologically impossible under the product conditions should not be treated as significant merely because it appears in a generic list, provided that conclusion is supported. The method may be qualitative or use a scoring matrix, but numbers do not replace judgement. Assumptions, uncertainty and the evidence behind the decision should be recorded.

Select and combine control measures

For each significant hazard the analysis identifies one or more control measures. Some operate through good hygiene practices or prerequisite programmes, such as supplier approval, sanitation, zoning or calibration. Others are specific process controls, such as a supported formulation, fermentation exposure, lethality schedule, cooling path, drying criterion or detection step. Cured meat often uses combinations: salt, nitrite, acidification, time, temperature, competitive flora, water activity and packaging may each contribute. The team must state what effect is claimed from each measure and whether the combination, sequence and operating range are supported. A measure that only inhibits growth should not be described as a kill step. A final endpoint should not be credited with controlling toxin already formed during an earlier deviation.

Determine the control route

After control measures are selected, the system determines how each will be managed. A significant hazard may be controlled through an effective general hygiene programme, an operational control or a critical control point, depending on the governing framework and the result of the analysis. Decision trees can assist but cannot replace professional judgement or create a safe answer from an incomplete hazard description. The rationale should explain why the selected route provides control and how failure will be detected. If no effective measure exists at the current or later step, the product or process must be changed. The team should also consider whether monitoring can identify the affected product and whether corrective action can be taken before release.

Document evidence, uncertainty and exclusions

Every significant decision should be traceable to evidence. The analysis records the source, its status, the conditions it covers and any gap between those conditions and the operation. Official guidance may identify hazards or accepted approaches; legislation may impose a limit; scientific work may support a control; establishment data may describe variability. None is used beyond its scope. Excluded hazards also require a defensible reason, particularly when the consequence is severe. Phrases such as not likely to occur should be supported by supplier controls, product characteristics, process evidence or other reliable information. Where uncertainty remains material, the conservative response is to obtain further support, modify the process or retain the hazard for control rather than assume that tradition or absence of previous incidents proves safety.

Review, learn and change

Hazard analysis is reassessed when the product, ingredient, supplier, equipment, process, packaging, storage, intended use or legal framework changes. New pathogen information, complaints, recalls, environmental results, repeated deviations and scientific developments also trigger review. The review asks whether new hazards have emerged, whether likelihood has changed, whether current controls remain capable, and whether monitoring still defines the affected scope. A new starter culture or casing may alter fermentation or drying; a larger diameter may change the time needed at the centre; a new slicer may add post-lethality exposure. Change control should update the hazard analysis, validation, procedures, training and records as one connected system. Editing the recipe alone leaves the safety rationale behind.